Clinical characteristics and management of PD-1/PD-L1 inhibitor-induced secondary adrenal insufficiency

Abstract

Graphical abstract

Abstract

Programmed cell death-1 (PD-1)/programmed death-ligand 1 (PD-L1) inhibitors significantly improve outcomes in multiple malignancies; however, they can induce immune-related endocrine adverse events, such as secondary adrenal insufficiency (SAI), a rare but potentially severe complication. We aimed to systematically analyze the clinical characteristics, changes in endocrine function, and treatment outcomes of SAI induced by PD-1/PD-L1 immune checkpoint inhibitors. This single-center retrospective study included 75 patients with tumors diagnosed with SAI following PD-1/PD-L1 inhibitor therapy. Clinical data were collected via electronic medical records, and clinical characteristics and prognosis data were analyzed using descriptive statistical methods. SAI occurred at any cycle during immunotherapy, with a median onset of 5.70 months. Clinically, fatigue was the most common symptom (80.0%), followed by loss of appetite (26.7%) and nausea (9.3%). Endocrine function assessments revealed that severe hypothalamic–pituitary–adrenal (HPA) axis impairment occurred in all patients; thyroid function abnormalities occurred in 34.7% of patients, with PD-1/PD-L1 inhibitor-related hypothyroidism accounting for 88.5%, of which 65.2% of hypothyroidism cases preceded SAI diagnosis. Central hypothyroidism (4.0%) and gonadal involvement (1.3%) were relatively rare; treatment outcome analysis showed that only 8.3% (1/12) of patients receiving high-dose glucocorticoid therapy recovered HPA axis function; 98.7% (74/75) of patients on long-term glucocorticoid replacement therapy tolerated subsequent anti-tumor treatment, and 59.5% maintained the regimen, including ICIs. These findings reveal the need to enhance endocrine monitoring during immunotherapy, prioritize HPA axis evaluation in patients with thyroid dysfunction, and promptly initiate standardized glucocorticoid replacement therapy upon SAI diagnosis to ensure continuity of anti-tumor treatment.

Introduction

Tumor immunotherapy has recently emerged as a significant milestone in cancer treatment, owing to its breakthrough clinical efficacy (1, 2). Immune checkpoint inhibitors (ICIs), including programmed cell death-1 (PD-1)/programmed death-ligand 1 (PD-L1) inhibitors, significantly extend survival in patients with lung cancer, liver cancer, melanoma, and other malignant tumors by relieving immune suppression in the tumor microenvironment (3, 4, 5, 6, 7, 8). With the continuous expansion of indications and optimization of combination treatment strategies, immunotherapy has progressed from second-line to first-line therapy and has even shown potential in neoadjuvant settings (9, 10, 11). However, while activating anti-tumor immune responses, immunotherapy may also overstimulate the autoimmune system, leading to immune-related adverse events (irAEs) (12, 13, 14). These events can affect multiple systems, with a high prevalence of endocrine involvement (15). Among endocrine irAEs, thyroid dysfunction (such as thyroiditis, hypothyroidism, and thyrotoxicosis) is the most common (16, 17, 18, 19), whereas hypophysitis and adrenal insufficiency, though less frequent, are more severe. Hypophysitis can lead to adrenocorticotropic hormone (ACTH) deficiency, resulting in secondary adrenal insufficiency (SAI), which often manifests as fatigue and hypotension (20). If not diagnosed and treated promptly, SAI can precipitate life-threatening adrenal crises, making early diagnosis and intervention critical. However, the non-specific clinical presentation of SAI often overlaps with symptoms of the tumor itself or chemotherapy side effects, leading to potential misdiagnosis or delayed treatment (20, 21, 22, 23, 24, 25). Although current guidelines recommend regular hormone level monitoring for patients receiving immunotherapy (26, 27, 28), prospective studies on ACTH deficiency-induced SAI are limited, particularly large-scale analyses of its incidence patterns, clinical characteristics, and long-term prognosis. Additionally, the impact of immunotherapy drug types (for example, differences between anti-CTLA-4 and anti-PD-1 antibodies) and combination regimens on SAI incidence requires further clarification. Therefore, we aimed to retrospectively analyze the clinical data of 75 patients with tumors who developed SAI following PD-1/PD-L1 immunotherapy to elucidate the incidence patterns, clinical characteristics, and management strategies of ACTH deficiency-induced SAI, providing evidence-based guidance for optimizing immunotherapy safety monitoring and standardizing management systems.

Materials and methods

Study population

This study used a single-center, retrospective cohort design, enrolling patients with solid tumors who received PD-1/PD-L1 inhibitor therapy at Harbin Medical University Cancer Hospital from January 2022 to January 2024.

The inclusion criteria were the following:

1. age ≥ 18 years,

2. histopathologically confirmed malignant tumors,

3. received PD-1/PD-L1 inhibitor monotherapy or combination therapy, and

4. met diagnostic criteria for SAI, defined as

5. clinical symptoms, such as fatigue, loss of appetite, and hypotension,

6. laboratory tests: morning (07:00–08:00) serum cortisol <83 nmol/L (3 μg/dL) and ACTH <1.6 pmol/L on two consecutive tests, and

7. imaging: head or pituitary magnetic resonance imaging (MRI) and abdominal computed tomography (CT) to rule out other causes of adrenal insufficiency.

The exclusion criteria were as follows:

1. history of hypothalamic–pituitary–adrenal (HPA) axis disorders, including prior pituitary or adrenal surgery or pituitary radiotherapy,

2. use of systemic glucocorticoids (prednisone >5 mg/day or equivalent) for more than two consecutive weeks within the 3 months prior to enrollment,

3. severe cardiac, hepatic, or renal insufficiency, and

4. incomplete clinical data.

Data collection

A standardized electronic data collection form was used to gather the following information via electronic medical records.

Baseline data

Demographic characteristics (age and sex), tumor characteristics (primary site, pathological type, and clinical stage (American Joint Committee on Cancer and 8th edition)), and treatment regimen (immunotherapy drugs, dosage, and administration cycle) were recorded.

Laboratory tests

The following tests were performed in our center’s laboratory: HPA axis: morning plasma ACTH and serum cortisol (chemiluminescence, Roche 801, Japan); thyroid hormones: thyroid-stimulating hormone (TSH), free thyroxine (FT4), and free triiodothyronine (FT3) (chemiluminescence, Roche 801); sex hormones: luteinizing hormone, follicle-stimulating hormone, estradiol, progesterone, and testosterone levels (chemiluminescence, Abbott i2000, USA); electrolytes: serum sodium and chloride levels (electrode method, AU5800); and others: fasting blood glucose levels and liver and kidney functions.

Imaging assessments

Pituitary MRI was performed using a 3.0T MRI scanner (Philips Achieva 3.0T, The Netherlands), including plain and dynamic contrast-enhanced sequences; adrenal CT was performed using a 64-slice spiral CT.

Statistical analyses

Data were analyzed using SPSS 26.0 software. For continuous data, normally distributed data are expressed as mean ± standard deviation; non-normally distributed data are expressed as median (interquartile range) (M (P25, P75)). Categorical data are expressed as frequency (percentage).

Research ethics

This was a retrospective analysis in which all treatment decisions were made based on clinical needs. The study strictly adhered to medical ethical principles, ensuring full protection of patient privacy. All enrolled patients provided signed informed consent, authorizing the use of their anonymized clinical data for medical research.

Results

Clinical manifestations

Among the 75 patients with tumors who developed ACTH deficiency-induced SAI following PD-1/PD-L1 inhibitor therapy, fatigue was the most common symptom, observed in 60 cases (80%). Patients typically presented with generalized weakness and significantly reduced endurance for daily activities. In severe cases, patients were unable to walk or perform activities. Gastrointestinal symptoms were the next most frequent symptoms, primarily manifesting as loss of appetite (n = 20, 26.7%), nausea (n = 7, 9.3%), vomiting (n = 3, 4%), and diarrhea (n = 2, 2.6%). Notably, some patients exhibited severe symptoms, such as frequent vomiting or diarrhea. One patient (1.3%) experienced severe hypoglycemia, presenting with palpitations and profuse sweating, with a venous blood glucose level of 2.0 mmol/L, which improved after intravenous glucose administration. Five patients (6.7%) had hypotension (blood pressure < 90/60 mmHg), with two (2.7%) exhibiting orthostatic hypotension, characterized by dizziness and transient vision loss upon changing from supine to standing position, resolving in approximately 1 min. Additionally, 12 patients had blood pressure below 95/65 mmHg but did not meet the criteria for hypotension, presenting only with generalized symptoms such as fatigue without specific hypotensive complaints (Table 1).

Table 1.

Clinical manifestations of 75 patients with PD-1/PD-L1 inhibitor-related SAI.

Clinical manifestation	Cases (n)	Percentage (%)	Detailed characteristics
Total cases	75
Fatigue	60	80	Generalized lethargy, reduced endurance for daily activities; in severe cases, inability to move or walk.
Loss of appetite	20	26.7	Decreased appetite.
Nausea	7	9.3	Nausea without vomiting.
Vomiting	3	4	Frequent vomiting in severe cases.
Diarrhea	2	2.7	Frequent diarrhea in severe cases.
Hypoglycemia symptoms	1	1.3	Palpitations, profuse sweating, venous blood glucose 2.0 mmol/L, improved after glucose infusion.
Hypotension symptoms	2	2.7	Dizziness, transient vision loss upon positional change from supine to standing, indicative of orthostatic hypotension.

PD-1, programmed cell death-1; PD-L1, programmed death-ligand 1; and SAI, secondary adrenal insufficiency.

Laboratory tests

Endocrine hormone test results

1. HPA: in accordance with the inclusion criteria, all patients demonstrated severe impairment of the ACTH–cortisol axis, with a median ACTH concentration of 0.22 pmol/L (Q1–Q3: 0.22–0.43 pmol/L). Notably, 61.3% (46 of 75) of patients exhibited levels below the detection threshold (<0.22 pmol/L).

2. Hypothalamic–pituitary–thyroid axis: thyroid dysfunction was observed in 26 patients, including three cases (4%) of central hypothyroidism (TSH 0.53 ± 0.56 mIU/L, FT4 8.32 ± 1.14 pg/mL, FT3 2.47 ± 0.50 ng/dL) and 23 cases (30.7%) of immunotherapy-related primary hypothyroidism (TSH 72.16 ± 25.44 mIU/L, FT4 5.40 ± 3.68 pg/mL, FT3 2.18 ± 1.05 ng/dL), with eight cases having TSH >100 μIU/mL, recorded as 100 μIU/mL for mean calculations. Among the 23 primary hypothyroidism cases, 11 progressed from thyrotoxicosis to hypothyroidism, 15 had hypothyroidism preceding cortisol deficiency, 5 had concurrent hypothyroidism and cortisol deficiency, 2 developed hypothyroidism after cortisol deficiency, and 1 developed hypothyroidism following concurrent thyrotoxicosis and cortisol deficiency. The remaining 49 patients (65.3%) had normal thyroid function.

3. Hypothalamic–pituitary–gonadal axis: only 20 patients (26.7%) underwent sex hormone testing, with 1 patient (1.3%) exhibiting central hypogonadism and 19 (25.4%) showing normal sex hormone levels (Tables 2 and 3).

Table 2.

Endocrine characteristics of 75 patients with PD-1/PD-L1 inhibitor-related SAI.

Evaluated axis	Abnormality category	Cases (percentage)	Test parameter (unit)	Test value	Remarks and reference range
HPA axis	Central cortisol deficiency	75 (100%)	ACTH (pmol/L)	0.22 (0.22–0.43)	↓↓ (normal: 1.6–13.9)
			Cortisol (nmol/L)	16.40 (8.23–27.30)	↓↓ (normal: 171–536)
HPT axis	Central hypothyroidism	3 (4%)	TSH (μIU/mL)	0.53 ± 0.56	↓ (normal: 0.27–4.2)
			FT4 (pmol/L)	8.32 ± 1.14	↓↓ (normal: 12–22)
			FT3(pmol/L)	2.47 ± 1.14	↓↓ (normal: 3.1–6.8)
	Immunotherapy-related hypothyroidism	23 (30.7%)	TSH (μIU/mL)	72.16 ± 25.44	↑↑ (includes 11 cases transitioning from thyrotoxicosis to hypothyroidism) (normal: 0.27–4.2)
			FT4 (pmol/L)	5.40 ± 3.68	↓↓ (normal: 12–22)
			FT3 (pmol/L)	2.18 ± 1.05	↓↓ (normal: 3.1–6.8)
	Normal thyroid function	49 (65.3%)	-	-	-
HPG axis	Not tested	55 (73.3%)	-	-	Only 20 cases (26.7%) tested
	Central hypogonadism	1 (1.3%)	FSH (mIU/mL)	1.23	Normal: 0.95–11.95
			LH (mIU/mL)	0.01	↓↓ (normal: 0.57–12.07)
			TESTO (nmol/L)	<0.45	↓↓ (normal: 4.94–32.01)
	Normal gonadal function	19 (25.4%)	-	-	-

HPA, hypothalamic–pituitary–adrenal; HPT, hypothalamic–pituitary–thyroid; HPG, hypothalamic–pituitary–gonadal; FSH, follicle-stimulating hormone; LH, luteinizing hormone; TESTO, testosterone; PD-1, programmed cell death-1; PD-L1, programmed death-ligand 1; and SAI, secondary adrenal insufficiency.

Table 3.

Temporal relationship between immunotherapy-related thyroid damage and cortisol deficiency (n = 23).

Temporal relationship	Cases	Percentage (%)
Hypothyroidism preceding cortisol deficiency	15	65.3
Hypothyroidism following cortisol deficiency	2	8.7
Hypothyroidism concurrent with cortisol deficiency	5	21.7
Thyrotoxicosis concurrent with cortisol deficiency	1	4.3
Total	23	100

Electrolyte and blood glucose abnormalities

Hyponatremia (serum sodium <135 mmol/L) was the most common electrolyte disorder, occurring in 40% (n = 30/75) of patients. By severity, it was mild (130–134 mmol/L) in 26.7% (n = 20/75), moderate (125–129 mmol/L) in 8% (n = 6/75), and severe (<125 mmol/L) in 5.3% (n = 4/75) of patients. Hypochloremia (serum chloride < 98 mmol/L) occurred in 25.3% (n = 19/75), with mild cases (90–97 mmol/L) in 20% (n = 15/75) and moderate cases (80–89 mmol/L) in 5.3% (n = 4/75), and no severe cases were observed. One patient (1.3%) experienced a severe hypoglycemic event (blood glucose 2.0 mmol/L) with typical autonomic symptoms (palpitations and profuse sweating). Additionally, six patients (7.9%) had blood glucose levels between 3.0 and 3.9 mmol/L, which, although not meeting hypoglycemia criteria and asymptomatic, suggested a potential risk of glucose metabolism abnormalities (Table 4).

Table 4.

Electrolyte and blood glucose levels in 75 patients with PD-1/PD-L1 inhibitor-related SAI.

Test parameter	Severity/category	Diagnostic criteria (mmol/L)	Cases (n)	Percentage (%)
Hyponatremia	Total	Serum sodium < 135	30	40
	Mild	130–134	20	26.7
	Moderate	125–129	6	8
	Severe	<125	4	5.3
Hypochloremia	Total	Serum chloride < 98	19	25.3
	Mild	90–97	15	20
	Moderate	80–89	4	5.3
	Severe	<80	0	0
Hypoglycemia	Total	Blood glucose < 2.8	1	1.3
	Others	Blood glucose 3.0–3.9	6	8

PD-1, programmed cell death-1; PD-L1, programmed death-ligand 1; and SAI, secondary adrenal insufficiency.

Imaging assessments

Pituitary MRI was performed in 12 patients (five plain scans; seven contrast-enhanced scans). Common findings included a normal or slightly enlarged pituitary volume with localized bulging at the superior or posterior margins. Signal characteristics included iso- or hypointense signals on T1-weighted imaging; iso-, hypo-, or hyperintense signals on T2-weighted imaging; and heterogeneous enhancement in the lesion area post-contrast. The pituitary stalk appeared normal, with no abnormal invasion of the optic chiasm or cavernous sinus.

Clinical characteristics of patients with adverse reactions

This study included 75 patients (55 males (73.3%), 20 females (26.7%], male-to-female ratio 2.75:1). The ages ranged from 39 to 79 years, with a mean of 60.83 ± 8.22 years. Age stratification was distributed as follows: <50 years (5 cases, 6.7%), 50–64 years (44 cases, 58.6%), and ≥65 years (26 cases, 34.7%).

Tumor types included lung cancer (n = 29, 38.7%), liver cancer (n = 10, 13.3%), gastric cancer (n = 8, 10.7%), esophageal cancer (n = 7, 9.3%), colorectal cancer (n = 5, 6.7%), pancreatic cancer (n = 3, 4.0%), renal cancer (n = 3, 4.0%), cholangiocarcinoma (n = 2, 2.7%), laryngeal cancer (n = 2, 2.7%), and one case each (1.3%) of tongue, nasopharyngeal, hypopharyngeal, thyroid, prostate, and penile cancers.

Immunotherapy regimen analysis showed that sintilimab (n = 26, 34.7%) and camrelizumab (n = 21, 28.0%) were the most commonly used drugs, accounting for 62.7% combined, followed by tislelizumab (n = 7, 9.3%), pembrolizumab (n = 7, 9.3%), toripalimab (n = 4, 5.3%), cadonilimab (n = 4, 5.3%), sugemalimab (n = 3, 4.0%), nivolumab (n = 2, 2.7%), and atezolizumab (n = 1, 1.3%). All patients were administered immunotherapy every 3 weeks (specific dosages in Table 5); 5 patients (6.7%) received immunotherapy alone, and 70 (93.3%) received immunotherapy combined with chemotherapy or targeted therapy (Table 5).

Table 5.

Clinical characteristics of 75 patients with PD-1/PD-L1 inhibitor-related SAI.

Clinical characteristics	Cases (n)	Percentage (%)	Remarks
Sex
Male	55	73.3
Female	20	26.7
Age (years)			60.83 ± 8.22
<50	5	6.7
50–65	44	58.6
>65	26	34.7
Tumor type
Lung cancer	29	38.7
Liver cancer	10	13.4
Gastric cancer	8	10.7
Esophageal cancer	7	9.3
Colorectal cancer	5	6.7
Pancreatic cancer	3	4
Renal cancer	3	4
Cholangiocarcinoma	2	2.7
Laryngeal cancer	2	2.7
Tongue cancer	1	1.3
Nasopharyngeal cancer	1	1.3
Hypopharyngeal cancer	1	1.3
Thyroid cancer	1	1.3
Prostate cancer	1	1.3
Penile cancer	1	1.3
Immunotherapy drug			Administration frequency (1 dose/3 weeks)
Sintilimab	26	34.7	200 mg per dose
Camrelizumab	21	28	200 mg per dose
Tislelizumab	7	9.3	200 mg per dose
Pembrolizumab	7	9.3	200 mg per dose
Toripalimab	4	5.3	240 mg per dose
Cadonilimab	4	5.3	500 mg per dose
Sugemalimab	3	4	300 mg per dose
Nivolumab	2	2.7	240 mg per dose
Atezolizumab	1	1.3	1,200 mg per dose
PD-1 inhibitor usage
Immunotherapy alone	5	6.7
Immunotherapy + chemotherapy/targeted therapy	70	93.3

PD-1, programmed cell death-1; PD-L1, programmed death-ligand 1; and SAI, secondary adrenal insufficiency.

Timing and outcomes of adverse reactions

Timing of adverse reactions

PD-1/PD-L1 immunotherapy-related SAI could occur at any treatment cycle (Fig. 1), with case distribution per cycle shown in Table 6, peaking during cycles 4–6. The onset time ranged from 0.7 to 16.8 months, with a median of 5.7 months (IQR, 3.70–10.30).

Figure 1.

Relationship between the incidence of SAI related to PD-1/PD-L1 inhibitor therapy and treatment cycles in 75 patients. PD-1, programmed cell death-1; PD-L1, programmed death-ligand 1; and SAI, secondary adrenal insufficiency.

Table 6.

Distribution of SAI complication onset by treatment cycle (n = 75).

Treatment cycle of complication onset	Cases	Percentage (%)
1	2	2.67
2	4	5.33
3	5	6.67
4	8	10.67
5	14	18.67
6	9	12.00
7	3	4.00
8	5	6.67
9	6	8.00
10	5	6.67
11	2	2.67
12	4	5.33
13	2	2.67
14	1	1.33
15	4	5.33
16	1	1.33
Total	75	100.00

SAI, secondary adrenal insufficiency.

Treatment and outcomes

Of the 75 patients, 12 (16%) received high-dose glucocorticoid anti-immune therapy (methylprednisolone 1–2 mg/kg body weight, reducing dosage by 20–25% every 3–5 days, for 4–6 weeks). One patient achieved normal ACTH and cortisol levels and discontinued immunotherapy, whereas the remaining 11 transitioned to long-term glucocorticoid replacement therapy. The other 63 patients (84%) directly received physiological-dose glucocorticoid replacement. All 74 patients on replacement therapy (11 from the high-dose group who did not recover and 63 from the direct replacement group) tolerated subsequent anti-tumor treatments well: 44 (59.5%) continued regimens including ICIs, whereas 30 (40.5%) switched to non-immunotherapy regimens (e.g. chemotherapy or targeted therapy) (Table 7).

Table 7.

Treatment groups and clinical outcomes of 75 patients with PD-1/PD-L1 inhibitor-related SAI.

Treatment group	Cases	Percentage (%)	Clinical outcome
A. High-dose glucocorticoid therapy	12	16.0
├─A1: hormone levels normalized	1	8.3*	Discontinued replacement therapy
└─A2: required long-term replacement	11	91.7*	Transitioned to the replacement therapy group
B. Initial replacement therapy	63	84.0	Continued glucocorticoid replacement
C. Total replacement therapy (A2 + B)	74	98.7
├─Continued ICI therapy	44	59.5†	Maintained anti-tumor regimen with ICIs
└─Switched to a non-immunotherapy regimen	30	40.5†	Used chemotherapy/targeted therapy
Overall tolerability	-	-	All 74 replacement therapy patients tolerated subsequent anti-tumor treatment

^*Percentage of the high-dose therapy group (n = 12).

^† Percentage of the total replacement therapy group (n = 74). ICIs, immune checkpoint inhibitors; PD-1, programmed cell death-1; PD-L1, programmed death-ligand 1; and SAI, secondary adrenal insufficiency.

Discussion

In this study, we retrospectively analyzed data of 75 patients with tumors who developed SAI following PD-1/PD-L1 ICI therapy to investigate its clinical characteristics, treatment, and prognosis. Our results revealed that PD-1/PD-L1 inhibitor-induced SAI can occur at any treatment cycle, with fatigue being the most prominent clinical manifestation, followed by nonspecific gastrointestinal symptoms, such as loss of appetite and nausea, and electrolyte imbalances, such as hyponatremia and hypochloremia. Fatigue, resulting from glucocorticoid deficiency-induced energy metabolism disorders, should be considered a primary warning sign of adrenal insufficiency during PD-1/PD-L1 inhibitor therapy. Gastrointestinal symptoms may be related to glucocorticoid deficiency-induced gastrointestinal motility and absorption dysfunction (29, 30). Persistent gastrointestinal symptoms, especially severe nausea, vomiting, or diarrhea, should prompt urgent evaluation of adrenal function. Notably, patients with tumors often present with fatigue, and those receiving combination chemotherapy usually experience gastrointestinal symptoms (such as loss of appetite, nausea, and vomiting). Therefore, during PD-1/PD-L1 inhibitor therapy, clinicians must remain vigilant and carefully differentiate new or persistent fatigue and unexplained gastrointestinal symptoms, proactively monitoring cortisol and ACTH levels to enable early SAI identification and prevent life-threatening adrenal crises. Additionally, rare but potentially fatal acute events, such as hypoglycemia and severe hypotension, warrant attention.

Regarding endocrine characteristics, all patients (100%) exhibited severe ACTH–cortisol axis suppression. In contrast, central damage to the TSH and gonadal axes was rare. This characteristic distribution suggests that PD-1/PD-L1 inhibitor-induced hypophysitis predominantly manifests as isolated ACTH deficiency-type anterior hypopituitarism. As corticotropin-releasing hormone stimulation testing was not performed, we could not precisely distinguish whether ACTH deficiency originated from the hypothalamus or pituitary. However, pituitary MRI findings generally supported a diagnosis of hypophysitis (e.g. heterogeneous signals, reduced enhancement, and posterior pituitary bulging). Notably, in this study, only 16% of patients underwent pituitary MRI examination, and the imaging characteristics differed from those in previous reports. Prior research has indicated that most patients with isolated ACTH deficiency exhibit no significant abnormalities on pituitary MRI. However, in this study, the majority of patients showed inflammatory changes (31). This discrepancy may be attributed to the timing of the examination: in this study, MRI was predominantly completed in the short term after patients presented with obvious clinical symptoms, such as severe fatigue and hyponatremia, making it easier to capture acute or subacute inflammatory changes in pituitary inflammation. However, previous studies may have primarily conducted examinations during the period of inflammation resolution or only mild functional abnormalities, resulting in insignificant imaging findings. Therefore, when immune-related pituitary inflammation is strongly suspected clinically, even if it manifests solely as isolated ACTH deficiency, it is advisable to perform pituitary MRI. Positive imaging results can provide important support for the diagnosis, although negative results do not rule out the diagnosis. Clinical and biochemical evaluations are key to the diagnosis. Future prospective large-scale studies and dynamic imaging follow-up are necessary to further elucidate the imaging evolution patterns and characteristics of this condition.

Although central hypothyroidism due to direct TSH axis involvement was rare (three cases), 30.7% of patients developed primary hypothyroidism due to PD-1/PD-L1 inhibitor-induced immune thyroiditis. Hypothyroidism could precede, coincide with, or follow cortisol deficiency, with precedence being the most common (65.4%), followed by concurrence. This temporal pattern suggests that the thyroid may be an earlier endocrine target of PD-1/PD-L1 inhibitors. Importantly, as thyroid damage is independent of the pituitary, patients with new-onset hypothyroidism following PD-1/PD-L1 therapy should have their HPA axis function evaluated prior to initiating thyroid hormone replacement to avoid precipitating potential adrenal crises. Morning cortisol and ACTH levels should be tested to rule out concurrent SAI. If cortisol levels are low, physiological-dose glucocorticoid replacement should be initiated immediately, with levothyroxine therapy starting only after adrenal function stabilization. If cortisol levels are normal, levothyroxine therapy may begin, but regular monitoring of morning cortisol and ACTH levels is necessary during subsequent immunotherapy to detect potential SAI.

Regarding gonadal function, the study’s limitation lies in incomplete sex hormone data (only 26.3% of patients were tested). Among the 20 patients tested, one exhibited central hypogonadism, whereas the others were normal. Despite the low incidence, proactive screening for patients with relevant risk symptoms is recommended for comprehensive pituitary function assessment.

For SAI treatment, we found significant differences in HPA axis function recovery based on glucocorticoid treatment strategies. Among the 12 patients who received high-dose glucocorticoid therapy, only one (8.3%) achieved HPA axis recovery (normal ACTH and cortisol levels), whereas 91.7% (11 cases) required long-term physiological-dose replacement. None of the 63 patients directly receiving physiological-dose replacement showed spontaneous HPA axis recovery, indicating that most immunotherapy-induced SAI cases involve irreversible HPA axis damage. Although high-dose therapy was attempted to reverse immune inflammation, its low recovery rate aligns with prior reports on immunotherapy-related hypophysitis prognosis (32, 33). Thus, early initiation of physiological-dose replacement may be a more practical choice (34, 35).

Notably, all 74 patients on replacement therapy (including 11 from the high-dose group who did not recover and 63 from the direct replacement group) tolerated subsequent anti-tumor treatments well. Of these, 59.5% (44/74) of patients continued the treatment regimen containing PD-1/PD-L1 inhibitors, whereas 40.5% (30/74) switched to other regimens (e.g. chemotherapy or targeted therapy). These outcomes highlight the dual value of glucocorticoid replacement: it prevents adrenal crises by maintaining homeostasis, ensuring safety for anti-tumor therapy, and enables nearly 60% of patients to continue PD-1/PD-L1 inhibitor therapy, confirming that SAI is not a contraindication for immunotherapy. With adequate replacement therapy, clinical decisions should balance tumor control needs with endocrine management goals, allowing most patients to benefit from continued immunotherapy. However, as a single-center retrospective analysis, these findings require validation through larger prospective studies.

Certain limitations of the study must be acknowledged. The majority of the PD-1/PD-L1 inhibitors used by patients in this cohort, such as sintilimab and carilizumab, have not received marketing approval in European and American countries. This may constrain the direct horizontal comparison of our results with the existing international literature, which predominantly focuses on approved drugs in Europe and America (such as pembrolizumab). However, these drugs are all immune checkpoint inhibitors that exert anti-tumor effects by blocking the PD-1/PD-L1 pathway. Their core mechanisms are highly consistent with those of approved drugs in Europe and America, making them comparable in terms of pathophysiology. Moreover, as these are standard therapies in China, our findings offer valuable real-world evidence for the management of ICI-related endocrinopathies, particularly within Asian populations.

In conclusion, this study demonstrated that PD-1/PD-L1 inhibitor-related SAI is characterized by pronounced fatigue and hyponatremia, frequently accompanied by immunotherapy-related thyroid dysfunction, which often precedes adrenal insufficiency. In terms of treatment, high-dose glucocorticoid therapy had a low HPA axis function recovery rate, whereas standardized physiological-dose glucocorticoid replacement therapy was safe and effective, enabling 98.7% of patients to successfully continue anti-tumor treatment, with 59.5% maintaining regimens that included ICIs. Based on these findings, we recommend enhanced endocrine function monitoring during immunotherapy, particularly prioritizing HPA axis evaluation in patients with thyroid dysfunction. Upon SAI diagnosis, standardized glucocorticoid replacement therapy should be promptly initiated to ensure the continuity of anti-tumor treatment.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work reported.

Funding

This work was supported by the Joint Fund Cultivation Project of Heilongjiang Provincial Natural Science Foundation (Grant No. PL2024H184), and Haiyan Foundation of Harbin Medical University Cancer Hospital (Grant No. JJMS2022-15).

Author contribution statement

NL conceptualized the research topic, designed the study protocol, organized and coordinated data collection, led data analysis and result validation, drafted the manuscript, and finalized revisions. HX was responsible for clinical data collection and organization and participated in statistical analysis. YN was responsible for data quality verification and participated in result analysis and discussion. TL guided the study design and implementation, supervised research quality and progress, and reviewed and approved the final manuscript.